|
Many analyses in environmental statistics involve displaying
spatial information, such as concentration of a pollutant in air or water, on a geographic
map. A problem which has had fairly limited attention is how many contour lines one should
use, and how uncertain they are. Some approaches to this problem will be illustrated using
a data set on tropospheric ozone in northeastern United States. Environmental statisticians
also frequently encounter compositional data, such as percentages of different chemical
species in particulate matter air pollution, or the proportion of pollution-tolerant,
pollution-intolerant and pollution-neutral species among bottom-dwelling aquatic animals.
An approach to model such data, and to display them, is illustrated on a problem of wood
smoke in Alaska, and (if time permits) benthic population in Delaware Bay.
|
|